Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 17 Oct 2004 03:42:11 +0200
From:      Volker <volker@vwsoft.com>
To:        Robert Watson <rwatson@FreeBSD.org>
Cc:        freebsd-current@freebsd.org
Subject:   Re: beta6/7 machine freeze
Message-ID:  <4171CDF3.6050901@vwsoft.com>
In-Reply-To: <Pine.NEB.3.96L.1041011114356.31040U-100000@fledge.watson.org>
References:  <Pine.NEB.3.96L.1041011114356.31040U-100000@fledge.watson.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Robert & list & RE,

sorry for delay. While I've done some hardware changes, I've been unable 
to check the i4b corrections earlier.

I've cvsup'ed to RELENG_5 again, recompiled using INVARIANTS and WITNESS 
enabled, setup ppp using i4b on the troubled machine and the machine 
didn't die again. I've closed the connection and re-setup the ppp link 
several times, also had a connection for more than two hours and the 
machine is still alive.

 From my point of view it looks like i4b is now healthy again and your 
patches made it.

I'm still using debug.mpsafenet=0 for the i4b code (may I remove it or 
is it still required for i4b?).

Before your i4b changes the machine died in the middle or at the end of 
a ppp connection. For me, this is a huge step forward (or backward into 
stability).

Now while the machine is running stable (no lock ups, no panics) for 
quite some days (the first time since BETA-4) I'm sure RELENG_5 made a 
huge step into the 5-STABLE direction. Two weeks ago I really had 
concerns for 5's stability but now I think 5.x will do a good job as 4.x 
did it.

I think you guys really did a great job! :)

Will remove INVARIANTS and WITNESS now and enjoy the speed of a good OS.

Thanks,

Volker



On 2004-10-11 17:44, Robert Watson wrote:

> In fact, this is almost certainly part of the problem.  I'd try setting
> debug.mpsafenet=0 in /boot/loader.conf and reboot as a first step t see if
> the problem goes away.
> 
> Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
> robert@fledge.watson.org      Principal Research Scientist, McAfee Research
> 
> On Mon, 11 Oct 2004, Robert Watson wrote:
> 
> 
>>BTW, since it sounds like this might be a networking related problem, you
>>might try setting debug.mpsafenet=0 in loader.conf to see if the problem
>>goes away -- that will put the Giant lock back over the network stack and
>>might close races.  i4b might not be MPSAFE, I'll go take a look.
>>
>>Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
>>robert@fledge.watson.org      Principal Research Scientist, McAfee Research
>>
>>
> 
> 

-- 
GPG/PGP fingerprint:
FF93 13A1 2477 B631 E953 06DF 4C49 ADD9 E4BF 79B1



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4171CDF3.6050901>